1
0

[HUDI-781] Add HoodieWriteableTestTable (#2040)

- Introduce HoodieWriteableTestTable for writing records into files
- Migrate writeParquetFiles() in HoodieClientTestUtils to HoodieWriteableTestTable
- Adopt HoodieWrittableTestTable for test cases in
  - ITTestRepairsCommand.java
  - TestHoodieIndex.java
  - TestHoodieKeyLocationFetchHandle.java
  - TestHoodieGlobalBloomIndex.java
  - TestHoodieBloomIndex.java
- Renamed HoodieTestTable and FileCreateUtils APIs
  - dataFile changed to baseFile
This commit is contained in:
Raymond Xu
2020-09-07 02:54:36 -07:00
committed by GitHub
parent 6537af2676
commit 83e39e2b17
13 changed files with 426 additions and 467 deletions

View File

@@ -21,6 +21,7 @@ package org.apache.hudi.common.testutils;
import org.apache.hudi.common.fs.FSUtils;
import org.apache.hudi.common.model.HoodieFileFormat;
import org.apache.hudi.common.model.HoodiePartitionMetadata;
import org.apache.hudi.common.model.IOType;
import org.apache.hudi.common.table.HoodieTableMetaClient;
import org.apache.hudi.common.table.timeline.HoodieTimeline;
@@ -40,6 +41,32 @@ import java.util.Map;
public class FileCreateUtils {
private static final String WRITE_TOKEN = "1-0-1";
public static String baseFileName(String instantTime, String fileId) {
return baseFileName(instantTime, fileId, HoodieFileFormat.PARQUET.getFileExtension());
}
public static String baseFileName(String instantTime, String fileId, String fileExtension) {
return FSUtils.makeDataFileName(instantTime, WRITE_TOKEN, fileId, fileExtension);
}
public static String logFileName(String instantTime, String fileId, int version) {
return logFileName(instantTime, fileId, version, HoodieFileFormat.HOODIE_LOG.getFileExtension());
}
public static String logFileName(String instantTime, String fileId, int version, String fileExtension) {
return FSUtils.makeLogFileName(fileId, fileExtension, instantTime, version, WRITE_TOKEN);
}
public static String markerFileName(String instantTime, String fileId, IOType ioType) {
return markerFileName(instantTime, fileId, ioType, HoodieFileFormat.PARQUET.getFileExtension());
}
public static String markerFileName(String instantTime, String fileId, IOType ioType, String fileExtension) {
return String.format("%s_%s_%s%s%s.%s", fileId, WRITE_TOKEN, instantTime, fileExtension, HoodieTableMetaClient.MARKER_EXTN, ioType);
}
private static void createMetaFile(String basePath, String instantTime, String suffix) throws IOException {
Path parentPath = Paths.get(basePath, HoodieTableMetaClient.METAFOLDER_NAME);
Files.createDirectories(parentPath);
@@ -73,45 +100,52 @@ public class FileCreateUtils {
createMetaFile(basePath, instantTime, HoodieTimeline.INFLIGHT_DELTA_COMMIT_EXTENSION);
}
public static void createDataFile(String basePath, String partitionPath, String instantTime, String fileId)
throws Exception {
createDataFile(basePath, partitionPath, instantTime, fileId, 0);
}
public static void createDataFile(String basePath, String partitionPath, String instantTime, String fileId, long length)
throws Exception {
public static void createPartitionMetaFile(String basePath, String partitionPath) throws IOException {
Path parentPath = Paths.get(basePath, partitionPath);
Files.createDirectories(parentPath);
Path dataFilePath = parentPath.resolve(FSUtils.makeDataFileName(instantTime, "1-0-1", fileId));
if (Files.notExists(dataFilePath)) {
Files.createFile(dataFilePath);
Path metaFilePath = parentPath.resolve(HoodiePartitionMetadata.HOODIE_PARTITION_METAFILE);
if (Files.notExists(metaFilePath)) {
Files.createFile(metaFilePath);
}
new RandomAccessFile(dataFilePath.toFile(), "rw").setLength(length);
}
public static void createLogFile(String basePath, String partitionPath, String baseInstantTime, String fileId, int version)
public static void createBaseFile(String basePath, String partitionPath, String instantTime, String fileId)
throws Exception {
createLogFile(basePath, partitionPath, baseInstantTime, fileId, version, 0);
createBaseFile(basePath, partitionPath, instantTime, fileId, 0);
}
public static void createLogFile(String basePath, String partitionPath, String baseInstantTime, String fileId, int version, int length)
public static void createBaseFile(String basePath, String partitionPath, String instantTime, String fileId, long length)
throws Exception {
Path parentPath = Paths.get(basePath, partitionPath);
Files.createDirectories(parentPath);
Path logFilePath = parentPath.resolve(FSUtils.makeLogFileName(fileId, HoodieFileFormat.HOODIE_LOG.getFileExtension(), baseInstantTime, version, "1-0-1"));
Path baseFilePath = parentPath.resolve(baseFileName(instantTime, fileId));
if (Files.notExists(baseFilePath)) {
Files.createFile(baseFilePath);
}
new RandomAccessFile(baseFilePath.toFile(), "rw").setLength(length);
}
public static void createLogFile(String basePath, String partitionPath, String instantTime, String fileId, int version)
throws Exception {
createLogFile(basePath, partitionPath, instantTime, fileId, version, 0);
}
public static void createLogFile(String basePath, String partitionPath, String instantTime, String fileId, int version, int length)
throws Exception {
Path parentPath = Paths.get(basePath, partitionPath);
Files.createDirectories(parentPath);
Path logFilePath = parentPath.resolve(logFileName(instantTime, fileId, version));
if (Files.notExists(logFilePath)) {
Files.createFile(logFilePath);
}
new RandomAccessFile(logFilePath.toFile(), "rw").setLength(length);
}
public static String createMarkerFile(String basePath, String partitionPath, String instantTime, String fileID, IOType ioType)
public static String createMarkerFile(String basePath, String partitionPath, String instantTime, String fileId, IOType ioType)
throws IOException {
Path folderPath = Paths.get(basePath, HoodieTableMetaClient.TEMPFOLDER_NAME, instantTime, partitionPath);
Files.createDirectories(folderPath);
String markerFileName = String.format("%s_%s_%s%s%s.%s", fileID, "1-0-1", instantTime,
HoodieFileFormat.PARQUET.getFileExtension(), HoodieTableMetaClient.MARKER_EXTN, ioType);
Path markerFilePath = folderPath.resolve(markerFileName);
Path parentPath = Paths.get(basePath, HoodieTableMetaClient.TEMPFOLDER_NAME, instantTime, partitionPath);
Files.createDirectories(parentPath);
Path markerFilePath = parentPath.resolve(markerFileName(instantTime, fileId, ioType));
if (Files.notExists(markerFilePath)) {
Files.createFile(markerFilePath);
}
@@ -119,11 +153,11 @@ public class FileCreateUtils {
}
public static long getTotalMarkerFileCount(String basePath, String partitionPath, String instantTime, IOType ioType) throws IOException {
Path markerDir = Paths.get(basePath, HoodieTableMetaClient.TEMPFOLDER_NAME, instantTime, partitionPath);
if (Files.notExists(markerDir)) {
Path parentPath = Paths.get(basePath, HoodieTableMetaClient.TEMPFOLDER_NAME, instantTime, partitionPath);
if (Files.notExists(parentPath)) {
return 0;
}
return Files.list(markerDir).filter(p -> p.getFileName().toString()
return Files.list(parentPath).filter(p -> p.getFileName().toString()
.endsWith(String.format("%s.%s", HoodieTableMetaClient.MARKER_EXTN, ioType))).count();
}

View File

@@ -19,8 +19,6 @@
package org.apache.hudi.common.testutils;
import org.apache.hudi.common.fs.FSUtils;
import org.apache.hudi.common.model.HoodieFileFormat;
import org.apache.hudi.common.model.IOType;
import org.apache.hudi.common.table.HoodieTableMetaClient;
import org.apache.hudi.common.util.ValidationUtils;
@@ -39,6 +37,7 @@ import java.util.Objects;
import java.util.UUID;
import java.util.stream.IntStream;
import static org.apache.hudi.common.testutils.FileCreateUtils.baseFileName;
import static org.apache.hudi.common.testutils.FileCreateUtils.createCommit;
import static org.apache.hudi.common.testutils.FileCreateUtils.createDeltaCommit;
import static org.apache.hudi.common.testutils.FileCreateUtils.createInflightCommit;
@@ -46,15 +45,16 @@ import static org.apache.hudi.common.testutils.FileCreateUtils.createInflightDel
import static org.apache.hudi.common.testutils.FileCreateUtils.createMarkerFile;
import static org.apache.hudi.common.testutils.FileCreateUtils.createRequestedCommit;
import static org.apache.hudi.common.testutils.FileCreateUtils.createRequestedDeltaCommit;
import static org.apache.hudi.common.testutils.FileCreateUtils.logFileName;
public class HoodieTestTable {
private final String basePath;
private final FileSystem fs;
private HoodieTableMetaClient metaClient;
private String currentInstantTime;
protected final String basePath;
protected final FileSystem fs;
protected HoodieTableMetaClient metaClient;
protected String currentInstantTime;
private HoodieTestTable(String basePath, FileSystem fs, HoodieTableMetaClient metaClient) {
protected HoodieTestTable(String basePath, FileSystem fs, HoodieTableMetaClient metaClient) {
ValidationUtils.checkArgument(Objects.equals(basePath, metaClient.getBasePath()));
ValidationUtils.checkArgument(Objects.equals(fs, metaClient.getRawFs()));
this.basePath = basePath;
@@ -124,6 +124,13 @@ public class HoodieTestTable {
return this;
}
public HoodieTestTable withPartitionMetaFiles(String... partitionPaths) throws IOException {
for (String partitionPath : partitionPaths) {
FileCreateUtils.createPartitionMetaFile(basePath, partitionPath);
}
return this;
}
public HoodieTestTable withMarkerFile(String partitionPath, IOType ioType) throws IOException {
return withMarkerFile(partitionPath, UUID.randomUUID().toString(), ioType);
}
@@ -150,19 +157,19 @@ public class HoodieTestTable {
*
* @return A {@link Map} of partition and its newly inserted file's id.
*/
public Map<String, String> withInserts(String... partitions) throws Exception {
public Map<String, String> withBaseFilesInPartitions(String... partitions) throws Exception {
Map<String, String> partitionFileIdMap = new HashMap<>();
for (String p : partitions) {
String fileId = UUID.randomUUID().toString();
FileCreateUtils.createDataFile(basePath, p, currentInstantTime, fileId);
FileCreateUtils.createBaseFile(basePath, p, currentInstantTime, fileId);
partitionFileIdMap.put(p, fileId);
}
return partitionFileIdMap;
}
public HoodieTestTable withUpdates(String partition, String... fileIds) throws Exception {
public HoodieTestTable withBaseFilesInPartition(String partition, String... fileIds) throws Exception {
for (String f : fileIds) {
FileCreateUtils.createDataFile(basePath, partition, currentInstantTime, f);
FileCreateUtils.createBaseFile(basePath, partition, currentInstantTime, f);
}
return this;
}
@@ -182,35 +189,37 @@ public class HoodieTestTable {
return this;
}
public boolean filesExist(Map<String, String> partitionAndFileId, String instantTime) {
public boolean baseFilesExist(Map<String, String> partitionAndFileId, String instantTime) {
return partitionAndFileId.entrySet().stream().allMatch(entry -> {
String partition = entry.getKey();
String fileId = entry.getValue();
return fileExists(partition, instantTime, fileId);
return baseFileExists(partition, instantTime, fileId);
});
}
public boolean filesExist(String partition, String instantTime, String... fileIds) {
return Arrays.stream(fileIds).allMatch(f -> fileExists(partition, instantTime, f));
public boolean baseFilesExist(String partition, String instantTime, String... fileIds) {
return Arrays.stream(fileIds).allMatch(f -> baseFileExists(partition, instantTime, f));
}
public boolean fileExists(String partition, String instantTime, String fileId) {
public boolean baseFileExists(String partition, String instantTime, String fileId) {
try {
return fs.exists(new Path(Paths.get(basePath, partition,
FSUtils.makeDataFileName(instantTime, "1-0-1", fileId)).toString()));
return fs.exists(new Path(Paths.get(basePath, partition, baseFileName(instantTime, fileId)).toString()));
} catch (IOException e) {
throw new HoodieTestTableException(e);
}
}
public String getBaseFileNameById(String fileId) {
return baseFileName(currentInstantTime, fileId);
}
public boolean logFilesExist(String partition, String instantTime, String fileId, int... versions) {
return Arrays.stream(versions).allMatch(v -> logFileExists(partition, instantTime, fileId, v));
}
public boolean logFileExists(String partition, String instantTime, String fileId, int version) {
try {
return fs.exists(new Path(Paths.get(basePath, partition,
FSUtils.makeLogFileName(fileId, HoodieFileFormat.HOODIE_LOG.getFileExtension(), instantTime, version, "1-0-1")).toString()));
return fs.exists(new Path(Paths.get(basePath, partition, logFileName(instantTime, fileId, version)).toString()));
} catch (IOException e) {
throw new HoodieTestTableException(e);
}